This study investigates the effectiveness of a Web search engine with newly added or improved features in Web search engine architecture. These features can be categorized into three groups: The impact of link quality and usage information on page importance calculation / the use of Turkish stemmer for indexing and query substitution / and, the use of thumbnails for Web search engine result visualization. As Web search engines have become the primary means for finding and accessing information on the Internet, the effectiveness of Web search engines should be evaluated on the idea of how effectively and efficiently they assist users achieve a query, which defines performance criteria rather than the pure precision and recall measures developed among basic information retrieval roles. In this thesis, we propose three distinguishing features to increase the efficiency of a Web search engine: The impact of link quality and usage information on page importance calculation outperforms classical hyperlink graph based methods notably, such as PageRank. The use of the Turkish stemmer for indexing and query substitution has remarkable improvements on Web relevance when used in a mixed framework with normal and stemmed forms. Finally, we have observed that users are able to find the most relevant results by using webpage thumbnails in the queries with decreased precision score values, despite their preferred search engine gazing behavior is much attributed.

The goal of this master's thesis is to develop stemming algorithm for czech language based on grammatical rules. You can find a description of stemming process and a comparsion of stemming algorithms in this project. The basics of czech grammar and Snowball language are also described here. The main part of this thesis concerns the implementation of the new czech stemming algorithm.

The main task of this Masters Thesis is to understand questions of information retrieval and text classifi cation. The main research is focused on the text data, the semantic dictionaries and especially the knowledges inferred from the Wikipedia. In this thesis is also described implementation of the querying system, which is based on achieved knowledges. Finally properties and possible improvements of the system are talked over.